摘要
基于电力系统中开关柜设备潜伏故障的检测需求,文中提出一种多传感器融合与机器学习相结合的智能故障检测方法,旨在提高故障诊断的准确性与实时性。针对传统故障检测方法在准确性和响应时间上的不足,采用红外热成像仪、激光诱导击穿光谱仪(LIBS)和UHF传感器,通过融合这三种传感器的数据,实时监测开关柜内温度、光谱和电磁波特征,结合支持向量机(SVM)和深度神经网络(DNN)算法建立故障识别模型。通过200次模拟局部放电实验,验证了该系统的有效性,系统的故障识别准确率达到99%,响应时间均值为9.12 s,满足实时监控的要求。实验结果表明,该智能故障检测系统能够有效提升开关柜故障诊断的准确性与可靠性,且对电力系统的安全运行具有重要保障作用。
Based on the detection needs of latent faults in switchgear equipment within the power system,this paper proposes an intelligent fault detection method combining multi⁃sensor fusion and machine learning,aiming to enhance the accuracy and real⁃time performance of fault diagnosis.Addressing the deficiencies in accuracy and response time of traditional fault detection methods,infrared thermal imaging,Laser⁃Induced Breakdown Spectroscopy(LIBS),and Ultra⁃High⁃Frequency(UHF)sensors are employed.By fusing data from these three sensors,real⁃time monitoring of temperature,spectral,and electromagnetic wave characteristics within the switchgear is achieved.A fault recognition model is established using Support Vector Machine(SVM)and Deep Neural Network(DNN)algorithms.The effectiveness of this system is verified through 200 simulated partial discharge experiments,achieving a fault recognition accuracy rate of 99%and a mean response time of 9.12 seconds,meeting the requirements for real⁃time monitoring.Experimental results demonstrate that this intelligent fault detection system can effectively improve the accuracy and reliability of switchgear fault diagnosis,and plays a crucial role in ensuring the safe operation of the power system.
作者
方江华
谢堪恒
贺平
黄蜓
石凌云
FANG Jianghua;XIE Kanheng;HE Ping;HUANG Ting;SHI Lingyun(Tongren Power Supply Bureau of Guizhou Power Grid Co.,Ltd.,Tongren 554300,China)
出处
《电子设计工程》
2025年第18期26-30,共5页
Electronic Design Engineering
基金
贵州电网科技项目(GZKJXM202322404)。
关键词
开关柜故障检测
局部放电
支持向量机
深度学习
switchgear fault detection
partial discharge
Support Vector Machine
deep learning
作者简介
方江华(1987-),男,贵州安顺人。研究方向:变电设备运维检修、智能变电站。
